Abstract
Background: Hospital Electronic Health Records (EHRs) increasingly capture health and functional deficits. We report outcomes for acute cardiac patients in relation to an automated frailty measure derived from these EHR data.
Methods: We conducted a retrospective observational cohort study of consecutive cardiology admissions aged ≥70 years between April 2016 and August 2020, to three hospitals across Edinburgh, Scotland. The Continuous Dynamic Evaluation of Frailty (CODE-f) is an automated score between 0 (no markers present) and 1 (all present) representing 12 deficits generated from 31 admission EHR data points. This includes measures of cognition, functional dependence, mobility and falls risk. The primary outcome was mortality at 1 year. The secondary outcome was days alive and out of hospital (‘home time’) in the year after discharge for hospital survivors. In a nested cohort of 318 consecutive patients, the Clinical Frailty Scale (CFS) was determined from manual EHR review blinded to CODE-f scores.
Results: 2,406 patients were included (mean 79±6 years old, 60% male). A CODE-f score could be generated in 2,158 (90%) patients, with a median score of 0.13 (IQR 0–0.33). There were 352 (15%) deaths by 1 year. Patients in the highest CODE-f quartile (>0.35) had three times greater risk of death at one year than in the lowest quartile after adjustment for age and sex (27% versus 9%, adjusted odds ratio 3.44, 95% CI 2.47–4.82, p<.001). 16% of patients from the highest CODE-f quartile lost>90 days home time in the year after discharge, compared to 6% in the lowest two quartiles (p<.001). CODE-f scores correlated moderately well with CFS (spearman’s r="0.50," 95% ci 0.41–0.58, p<0.001).
Conclusion: An automated EHR measure can identify older adults at risk of death and poorer recovery after acute cardiac illness. This could inform treatment decisions future care planning.
Funding: Chief Scientist Office (pcl />18/05)